There is known a conventional technique of: calculating processing times of services by receiving request messages for ordering execution of the services and response messages indicating responses to the respective request messages; and changing the number of concurrent processes for services such that the processing time of each service can fall within a predetermined time range (see, for instance, Japanese Laid-open Patent Publication No. 2014-38578).